STRUCTURE
IN
PLANT
OF
CORTICAL
MICROTUBULE
ARRAYS
CELLS
A. R. HARDHAM and B. E. S. GUNNING
From the Department of DevelopmentalBiology,Research Schoolof BiologicalSciences,Australian
National University,Canberra, A.C.T. 2601, Australia
ABSTRACT
Serial sectioning was used to track the position and measure the lengths of
cortical microtubules in glutaraldehyde-osmium tetroxide-fixed root tip cells.
Microtubules lying against the longitudinal walls during interphase, those overlying developing xylem thickenings, and those in pre-prophase bands are oriented
circumferentially but on average are only about one-eighth of the cell circumference in length, i.e., 2-4 /~m. The arrays consist of overlapping component
microtubules, interconnected by cross bridges where they are grouped and also
connected to the plasma membrane. Microtubule lengths vary greatly in any
given array, but the probability that any pass fight around the cell is extremely
low. The majority of the microtubule terminations lie in statistically random
positions in the arrays, but nonrandomness in the form of groups of terminations
and terminations in short lines parallel to the axis of cell elongation has been
observed. Low temperature induces microtubule shortening and increases the
frequency of C-shaped terminations over the 1.7% found under normal conditions; colchicine and high pressures produce abnormally large proportions of
very short microtubules amongst those that survive the treatments. Deuterium
oxide (DzO) treatment probably induces the formation of additional microtubules
as distinct from increasing the length of those already present. The distribution
of C-shaped terminations provides evidence for at least local polarity in the
arrays. The validity of the findings is discussed, along with implications for the
development, maintenance, and orientation of the arrays and their possible
relationship to the orientation of cellulose deposition.
KEY WORDS microtubules
pre-prophase bands
xylem
plant cell cortex
serial sectioning
The discovery of cortical microtubules in plant
cells followed the introduction of glutaraldehyde
as a fixative for fine structural studies. Ledbetter
and Porter (28) described them as being "arranged circumferentially [against the lateral walls
of root tip cells] . . . like hundreds of hoops
around the cell," parallel to microfibrils of cellulose in the cell wall. While admitting that individ14
ual microtubules could be followed for only short
distances, Newcomb (34) stated that "In some
cases they run circumferentially around the cell in
the cytoplasm near the plasmalemma, probably
forming complete rings or hoops." Green et al.
(13) also considered that cortical "microtubules
are arranged in a very loose spiral or are present
as rings." Srivastava and Singh's description (53)
of microtubules overlying developing xylem thickenings is that they too "run like hoops around the
cell," and Hepler and Palevitz (21) refer to those
J. CELL BIOLOGY9 The RockefellerUniversity Press-0021-9525/78/0401-14 $1.00
in a differentiating sieve e l e m e n t as being "arr a n g e d hooptike a r o u n d the cell." Even m o r e
recently, Pickett-Heaps (41) envisaged that "cytoplasmic microtubules normally encircle the cell
transversely to the long axis in b o t h higher plants
a n d many algae." Schnepf et al. also implied that
the cortical microtubules are long in suggesting
"'unusually fast elongation of the microtubules
[along a newly f o r m e d wall] so that they soon
a p p e a r on the opposite wall . . . " (48).
T h e r e is, however, growing evidence that mic r o t u b u l a r arrays can contain overlapping r a t h e r
than full-length microtubules (1 l , 18, 20, 23, 26,
32, 33, 59). In the case of the plant cell cortex,
this might be of especial i m p o r t a n c e . It has bec o m e evident that the c o n g r u e n t alignment of
cortical microtubules a n d wall microfibrils is widespread during primary and secondary wall deposition (21), and any a t t e m p t to specify functional
aspects of this co-orientation is crucially dependent upon knowledge of the actual three-dimensional a r r a n g e m e n t of the microtubules. A t least
one hypothesis views the microtubules as providing rigid guide tracks for the m o v e m e n t of cellulose synthetase complexes (19). Yet, there is in
fact no compelling evidence that cortical microtubules are long h o o p s or helices. W a r n i n g s given
by the original discoverers that they could equally
be in the form of overlapping arcs (27, 29) have
gone largely u n h e e d e d . T h e main aim of the
present work was to provide basic information on
the g e o m e t r y of cortical microtubule arrays in
plant cells, this being f u n d a m e n t a l to an understanding of how they are f o r m e d and m a i n t a i n e d ,
and how they might function. Preliminary data
have b e e n published (16, 17).
MATERIALS
AND
METHODS
Plant Tissue and Routine Fixation
Root tips of Azolla pinnata R.Br. (comprehensive
descriptions in preparation), Impatiens balsamina L. and
Zea mays L. were fixed in 2.5% glutaraldehyde in
0.025 M phosphate buffer, pH 7, for 2-17 h at room
temperature, rinsed in buffer, and postfixed in 2%
osmium tetroxide in the same buffer for 2 h at room
temperature. The tissue was dehydrated in graded acetone solutions and embedded in Spurr's resin (52).
Variations from this standard procedure are described
in Table I and appropriate parts of Results.
Experimental Treatments
For low temperature treatments, intact Azolla pinnata
plants were transferred to culture medium in beakers
surrounded by ice in a cold room, for either 15 rain or 4
h. The temperature of the water in which the roots were
immersed varied within the range 0~176 in different
experiments (Table VI). Roots were fixed for 1 h at this
temperature and then for a further hour at room temperature. In a recovery experiment, roots were transferred
to room temperature culture medium for 15 rain before
fixation at room temperature. The rate of cooling and
warming in the actual root tissue was not measured, but
since the diameter of the root at the level that was
analyzed is only ~150 p,m, it is assumed to be rapid.
Treatments with colchicine in the dark at 5 • 10-a M
lasted 2, 3, 4. or 5 h, as well as 2-h treatment followed
by a 1-h recovery. Deuterium oxide (D20) was used at
99.8% purity for 5 or 18 h before fixation (in glutaraldehyde dissolved in D20). Pressures of 6,000 lb/in2 (for
30 min), 14,000, or 16,000 lb/in2 (for 15 min) were
applied to intact Azolla pinnata plants, using a French
pressure cell. The pressure was released smoothly but
quickly, and roots were immersed in fixative within 30 s
of the onset of decompression.
Serial Sectioning
Ribbons of sections cut from longitudinally oriented
roots were collected on parlodion-coated grids with 2mm slots and stained with saturated uranyl acetate in
50% ethanol for 30 min, then in lead citrate for 10 min
(56), and viewed in a Hitachi H500 electron microscope.
Suitable preparations were photographed at x 30,000
and printed to a final magnification of 75,000.
Tracing and Mapping
Microtubule profiles were traced from the micrographs of each serial section onto cellulose acetate
sheets. Individual microtubules were followed by placing
the tracing of one section over the micrograph of the
adjacent section in the sequence. When the entire sequence had been analyzed in this way and the positions
of terminations and the paths of the microtubules determined, the position of each microtubule profile on the
tracings was plotted on graph paper so as to build up a
map of the tubules in the plane of the array (e.g. Figs.
4-6). Terminations were arbitrarily taken to occur halfway through the section in which they had been detected. Where possible, a microtubule which continued
right through the sequence was used as a reference and
each successive section was aligned with respect to it.
Spatial separation of microtubules in the direction from
the wall to the center of the cell cannot be depicted on
these maps, and some microtubules which appear to lie
very close together may, in fact, be some distance apart.
In preparing the maps of some very dense arrays, it was
necessary for the sake of clarity to displace the positions
of a few microtubules just enough to show their existence.
Estimation o f Section Thickness
The construction of the microtubule maps described
above and the calculation of microtubule lengths is
HARDttAM AND GuNr~1r~
CorticalMicrotubules in Plants
15
dependent on knowledge of the section thickness. Ideally, the thickness of every section should be measured,
so that variation such as is evident in Fig. 3 could be
taken into account. We have adopted a compromise
procedure, calibrating our ultramicrotome by estimating
the depth to which ribbons of ultrathin sections of
various interference colors cut into the block. Details
are given in reference 15. An average section increment
of 70 nm was estimated for the ribbons of serial sections
used in the present work (this is a slight revision of the
value given in reference 17).
Derivation of the Average
Microtubule Length
Consider the microtubules to be arcs lying adjacent
to the walls of a cylindrical cell, perimeter A, as viewed
in transverse section in Fig. 1. The average length of the
microtubule arcs, L, is assumed to be less t h a n A . The
cell is sectioned longitudinally and the total thickness of
the combined serial sections is the sample arc, a. The
lengths are assumed to be independent of position and
of each other, the midpoints of the microtubules forming
a poisson process of intensity, X. In this situation, the
expected number of terminations (7) occurring in the
sample arc is given by T = 2ha. The expected number
of microtubules (N) passing through a fixed point is N =
hL, and hence the average microtubule length can be
calculated from the relationship L = 2Na/T.An intuitive
derivation is given in Results, and the validity of the
formula is easily checked using imaginary arrays drawn
on graph paper. The use of sections of finite thickness
introduces an error which can be avoided if, when
counting the number of profiles of microtubules per
section in order to derive N (the average number per
section), one unit is added for each tubule that is not
terminating, and one-half unit for each tubule that
terminates in that given section. Thus, it is assumed that
the terminations always lie at the midplane of the
sections.
Analysis of the Spatial Distribution
of Terminations
Objective methods for detecting possible nonrandomness in the distribution of terminations were applied. A
multiple X2 test was used to compare the observed
distribution of terminations amongst the sections in each
sequence with that predicted by a poisson distribution
calculated using the average number of terminations per
section as its mean. The significance of calculated values
of X2 was determined from the table given in reference
10. All of the arrays described in Results as "random"
had probabilities of correspondence with poisson distributions > P = 0.05 (mostly >>), and of those described
as "nonrandom" the probabilities of correspondence
were ~ P = 0.01 - 0.02 (mostly "~:P = 0.01). This
analysis, however, indicates deviation from randomness
only if the nonrandomly distributed terminations lie in a
pattern that is detectable by the sampling procedure,
i.e., by the cutting of sections in a particular plane.
Evidence for nonrandomness due to dumping of terminations over several successive sections, or to lines of
terminations not lying in the plane of the section, may
be obtained by applying a pattern analysis which measures the variability or spread of a distribution of points
(14, 24). In this procedure, a microtubule map is
subdivided by a grid to give convenient grid units or
"blocks," and the number of terminations in each block
is counted. If their distribution is random the variance is
equal to the mean, whereas larger values of variance
indicate clustering, and lower values, regular dispersion.
The variance of the population is calculated at different
unit block sizes. The ratio of variance to mean is plotted
against size, and confidence bands (55) are applied to
determine sizes at which the distribution deviates significantly (P = 0.05) from randomness.
RESULTS
Microtubule Terminations in Longitudinal
and Transverse Views
FIGURE 1 Microtubules, length L, lie adjacent to the
perimeter of a cell, circumference, A. The sample arc,
a, is equal to the thickness of the combined serial
sections.
16
THE JOURNAL OF CELL BIOLOGY " VOLUME
77,
If an ultrathin section is 75 nm (i.e. three
microtubule diameters) thick, any microtubule
whose long axis deviates from the plane of section
by as little as 4 ~ will only have, at most, 1 /zm of
its length included within the section. T h e zone
over which the microtubule leaves (or enters) the
surface of the section will show a gradual loss of
image of the tubule; Fig. 2 a and b, which show
two adjacent serial sections t h r o u g h a pre-prophase b a n d of microtubules, contain m a n y examples. In theory, it should be possible to follow
microtubules in longitudinal view from one section
1978
FmURE 2 Two adjacent serial sections through a pre-prophase band of microtubules in an Azolla root
tip cell illustrate the difficulty of tracking microtubules in longitudinal section. Microtubules that have
been sectioned transversely can, however, be followed from one section into the next (see boxes).
Periodic bridging between adjacent longitudinally sectioned microtubules is evident in Fig. 2a (arrows).
x 100,000.
to another until terminations are found. In practice, it is very difficult to do so, especially in
dense arrays of microtubules. However, where
the microtubules lie at right angles to the plane of
section, the relative positions of the circular transverse sectional profiles allow the presence or
absence of a particular microtubule to be determined with confidence (Fig. 2a and b).
Examples of microtubule terminations can be
seen by following the numbered microtubules in
Fig. 3. 18 of the 42 microtubules that are present
in this part of the sequence end within the sections
shown, the terminations occurring throughout the
array. The sequence includes one short complete
microtubule which begins in Fig. 3 c and ends in
Fig. 3j.
No electron-dense material was consistently
seen at termination sites, but 1.7% of the terminations were found to possess a C-shaped profile
in transverse section (see Table V), as exemplified
by microtubule n u m b e r 2 in Fig. 3 a. This confor-
mation usually occurs in the last one or two
sections but has been seen for up to 14 sections
(1 /.Lm) from the end of the microtubule. On two
occasions, microtubules were found to possess Cshaped profiles which were isolated from the
terminations by complete, circular profiles. Sometimes the arms of the "C" are much reduced,
leaving only a small arc of the original microtubule
profile.
The 12 C-shaped terminations (1.7% of a total
of 706) showed nonrandom distribution patterns
within the arrays. They were present in just three
of the sequences of serial sections (all from different cells), one having two and the others four and
six, respectively. C-shapes were never found at
both ends of an individual microtubule, and if
one mentally symbolizes a C-shape as an arrowhead at one end of a microtubule, then, in the
array that had two C-shapes, both "arrows"
pointed in the same direction. In the array with
four C-shapes, all four pointed in the same direc-
HARDHAM AND GUNNING Cortical Microtubules in Plants
17
FIGURE 3 Eleven serial sections (a-k) through a pre-prophase band of microtubules in a root tip cell of
Azolla. 18 of the 42 microtubules in this part of the array terminate within these sections. Examples of
microtubule terminations can he seen by following the selected numbered microtubules: microtubules
numbered 1 and 3 make their last appearances in Fig. 3 i and h, respectively; rnicrotubule 5 makes its first
appearance in Fig. 3 d. A C-shaped termination is illustrated by microtubule 2 in Fig. 3 a. A microtubule
containing a zone in which the staining is much fainter than in adjacent regions is arrowed in Fig. 3 d .
The entire length of microtubule 4 is encompassed within these 11 sections. Microtubules passing at a
greater angle than the majority in the array, e.g. those in Fig. 3c (arrowheads), were not included in the
mapping or calculations. Note bridges between microtubules and the plasma membrane (horizontal
arrowheads). • 106~000.
tion, and the same applied to the array with six
C-shapes. Within a given array, the microtubules
with C-shaped terminations evidently share a common directionality, at least within the portions of
the arrays encompassed in the sequences of serial
sections.
l
;o
20
I
Bridges
Several examples of cross bridges between adjacent microtubules and also between microtubules and the plasma membrane are illustrated in
Figs. 2 and 3. The bridging between the longitudinally sectioned microtubules in Fig. 2 a exhibits
a herringbone pattern with a periodicity approximating to the diameter of the microtubules themselves. Bridges attached to transversely sectioned
microtubules are not readily distinguished from
other material that may be lying close to the
microtubules. However, relatively clear images of
putative bridges were seen in all three categories
of cortical array: in pre-prophase bands (Figs. 2
and 3), in interphase arrays, and in arrays of
microtubules overlying developing xylem thickenings. They are present on short as well as long
microtubules. In a number of cases, a cross bridge
either to the plasma membrane or to an adjacent
microtubule occurred on the terminating profile
of a microtubule. On occasions, two bridges connected a single microtubule profile to the plasma
membrane. Cross bridges between microtubules
in pre-prophase bands and vesicles were occasionally seen.
Treatments which are known to stabilize microtubules or which can cause their depolymerization
were used in parts of this study (see later): following all such treatments, the appearance and distribution of the microtubule cross bridges were
similar to those described above in control tissues.
Microtubule Maps
It is evident from the microtubule maps (Figs.
4-6) that, in all three types of cortical microtubule
array described below, some microtubules continue right through the sequence, others have one
end within the sequence, and a few have both
ends in the sequence, that is, the complete microtubules lie within the thickness of tissue sectioned.
Many neighboring microtubules maintain the
same relative spacing within small bundles over
long distances (Fig. 6), but some show undulations. Some (usually lying deeper in the cytoplasm) pass at an angle to the remainder, and
Number of section ak~g sequence
FIGURE 4 Two reconstructions of arrays of microtubules in developing xylem elements. In Fig. 4A the
microtubules were grouped over two clearly discernible
wall thickenings at the beginnings of the sequence;
however, by section 8 the thickenings were no longer
visible. The reconstruction shows that in this region the
microtubules begin to fan out, becoming more evenly
distributed over the plasma membrane. Fig. 4 B depicts
an array overlying two thickenings in material that had
been treated with D~O for 5 h. The thickenings were
more developed than those in Fig. 4 A and continued
throughout the sequence of sections. In these maps, as
in all others except Fig. 6 A, the arrays are viewed as
from the cell wall looking into the cell through the
plasma membrane.
some leave one bundle and slant across to join
another bundle. There is a tendency for isolated
tubules to have atypical orientations. Microtubules lying at large angles to the majority were
occasionally seen in the sequences of sections,
e.g. Fig, 3, but were not included in the maps or
calculations.
DIFFERENTIATING
XYLEM
ELEMENTS:
An early correlate of the initiation of wall thickenings in developing xylem is the appearance of
groups of microtubules along the longitudinal
walls (39, 40). In studying the development of
xylem elements in Azolla pinnata roots, it was
often observed that microtubules were grouped on
one side of the cell while they were distributed
more evenly along the opposite wall. One reconstruction of a microtubule array in a xylem element followed the convergence from a dispersed
array to a pattern of groups, visible wall thickening commencing where the groups are clearly established (Fig. 4A). In reconstructions of other
elements, the microtubules remained grouped
over thickenings throughout the sequences.
PRE-PROPHASE
BANDS;
Pre-prophase
HARDHAM AND GUNNING CorticalMicrotubules in Plants
19
bands, the numbers of microtubules on opposite
sides of an interphase cell are not always equal.
While in most sequences of serial sections the
numbers of microtubule profiles per section remained relatively constant (e.g. Fig. 6 A), a few
cases were found where they slowly changed (e.g.
Fig. 2 in reference 17). The array shown in Fig.
6B includes a very abrupt change. In this case,
11 of the 13 microtubules in section 24 terminated
and only two continued into section 22. As the
sequence continued, the number of microtubules
slowly increased again. Fig. 6 A contains a less
obvious line of terminations at the same part of
the sequence, the two arrays in Fig. 6 lying back
to back, <1 /zm apart in adjacent interphase
cells.
Analysis of the Distributions of
Microtubule Terminations
Number of section along sequence
1pm
FIGURE 5 A reconstruction of part of a pre-prophase
band of microtubules. The dimension normal to the cell
surface is not apparent, but only 57-72% of the microtubules were adjacent to the plasma membrane. Most
microtubules lie parallel to the plasma membrane, and
are not interwoven.
bands of microtubules (41) occur throughout
Azolla root tips, encircling the cells and anticipating the plane of every category of cell division
(details in preparation). The number of microtubule profiles in a band varies widely, from 30 to
> 100. Where the number is small, most lie close
to the plasma membrane; where it is large, the
width of the band remains - 2 /xm and the microtubules become stacked. The number can change
(Fig. 5), sometimes markedly (Fig. 8), within a
short sequence of sections. In the five pre-prophase bands that were serially sectioned, between
48 and 74% of the microtubule terminations
occurred near the plasma membrane. In general,
this percentage approximates to the percentage of
microtubules that lie adjacent to the plasma membrane. Further analysis of the positions of terminations in pre-prophase arrays shows that, for
90% of the microtubules that were completely
included in the sequences of sections, both terminations of any given microtubule lay at the same
distance from the plasma membrane.
INTERPHASE ARRAYS: As in differentiating xylem elements, and in pre-prophase
20
The observed distribution of microtubule terminations throughout the sequences of sections
was compared to theoretical distributions calculated for each sequence from poisson series in
which the mean is the number of terminations
per section. In the majority of cases obtained
from untreated root tips, the observed dispersion
of terminations did not differ significantly from
the poisson distributions, but in three interphase
arrays and one pre-prophase band deviation from
randomness occurred, due to the presence of
either (a) clusters of terminations where most of
the members of a parallel group ended within a
few sections (as in two interphase arrays, e.g.
Fig. 2 in reference 17) or (b) linear arrangements
of terminations (as seen in the interphase array in
Fig. 6B and in the pre-prophase band). Other
linear arrangements of terminations were seen,
but since the line did not coincide exactly with
the plane of sectioning, the overall agreement
with the poisson distribution was retained. Similarly, some microtubule maps gave visual impressions of clustering but because the terminations,
although clustered, were dispersed amongst several sections, the overall distribution did not deviate significantly from the poisson series.
A pattern analysis which can show whether a
set of points is randomly distributed, clustered, or
regularly dispersed was also used. When applied
to the microtubule arrays, significant clustering of
terminations was found to occur in all of the
sequences which had, and three sequences which
had not, deviated from the poisson distribution.
Thus, in seven of the 16 arrays analyzed in this
THE JOURNAL OF CELL BIOLOGY-VoLuME 77, 1978
1 1 1 ~ 1 1 1 1 1 1 1
I I !
2oI I I ,,
I- I I I
3o
~ I I I I I I I
I { I t
t t
fi
I I I I
I I
5o
6o
,,,
I I
I I I
?o
I~L
a
i
lu
II
to
ii
lJl
i
l l l l t t
I~
20
t
1l[
I$ I
I11
I
30
I tl
I I I t [ l t t
40
1]11
I I I I I
10
II]
1 t t l l
60
tl
I1
I t 1 I ~ 1
70
I t l t
It
II
II
I11
III
J
1
50
I I I ~ 1
20
J
3o
40
50
6o
70
Number of section along sequence
llxn
FIGURE 6 Scale representations of two arrays of cortical microtubules against the longitudinal cell walls
at interphase. The view in Fig. 6 A is as from the center of the cell, looking outwards. Fig. 6 B represents
the cortex of the neighboring cell, viewed by looking on through Fig. 6 A and the intervening wall. The
two arrays can be superimposed in register by using the left-hand frame and the asterisks as reference
marks. The axis in which the cells were elongating runs up and down the page. Azollapinnata.
part of the work, the distribution of microtubule
terminations showed some form of nonrandomhess.
The Average Length o f Microtubules
The sequences of serial sections and the maps
drawn from them show that the observed microtubule lengths range widely from microtubules
which are recognizable in only one section up to
the longest microtubule so far detected, which
was still continuing after 84 sections (5.9 /zm).
Despite the variation in length, semiquantitation
is possible. There are two terminations for each
microtubule in an array when the total thickness
of the combined serial sections (a) just matches
the average microtubule length (L). Sequences of
sections that are shorter or longer than this will
contain correspondingly fewer or more terminations:
no. of terminations in sequence (7)
2 • average number of microtubules per section (N)
=
a
_
L"
A more rigorous derivation of this formula is set
out in Materials and Methods. Strictly, it should
only be used where the distribution of terminations conforms to a poisson distribution. The data
HARDHAM AND GUNNrN6 CorticalMicrotubulesin Plants
21
have, however, been pooled because the nonrandom arrays were not obviously different from the
random in terms of calculated average length. In
calculating microtubule lengths, it was assumed
that all microtubules lay at right angles to the
plane of section and that the cell cortex was flat
rather than curved: both assumptions lead to
slight underestimation of true lengths.
STANDARD
FIXATION
PROCEDURE:
The
data from the sequences of serial sections through
arrays of cortical microtubules in Azolla pinnata
root tip cells and the calculated average lengths
are shown in Table I. There is considerable deviation from the mean in each category, but the
average lengths, in general, are - 2 - 4 /zm, irrespective of the type of array. Values of 3-4 t~m
were obtained for Zea mays and 5-6 /zm for
Impatiens balsamina (Table II). In all three plants,
the average microtubule length was approximately
one-eighth of the circumference of the cells that
were examined.
ALTERNATIVE
FIXATION
PROCE-
The duration of glutaraldehyde prefixation (see Table I) was varied when preparing the
material described in the preceding section, and it
was clear that these variations had no major effect
on the results. Azolla root tips were also fixed in
2.5% glutaraldehyde in a polymerization medium
containing 1 mM guanosine triphosphate (GTP),
1 mM magnesium sulphate, 2 mM ethylene glycol-bis(fl-aminoethyl ether)N,N'-tetraacetic acid
and 100 mM piperazine-N,N'-bis[2-ethane sulphonic acid] buffer, pH 6.9. In other experiments
the GTP was omitted. The general ultrastructure
of the cells, and more specifically, of the microtubule arrays, did not differ from that seen after
conventional fixation in phosphate-buffered glutaraldehyde. A reconstruction of an interphase
array in tissue fixed in the complete polymerization medium is shown in Fig. 7. The average
microtubule lengths (Table III) were slightly
longer than in conventionally prepared roots, but
the difference is not significant. Fixation of roots
in polymerization medium lacking GTP made no
significant difference with respect to the average
microtubule length. None of the 181 terminations
observed after fixation in polymerization medium
was C-shaped (Table V).
A root fixed in polymerization medium provided another example of abrupt changes in the
numbers of microtubules per section due to the
presence of lines of terminations. The 16-section
sequence cut through an initially highly stacked
DURES:
22
TABLE I
Average Lengths of Cortical Microtubules in Azolla
Root Tip Cells
Type of
microtubule
array
No. of
sections
in sequence
(x0.07
= a, v,m)
Average
no. of
microtubules
per
section
(N)
No. of
terminations
in sequence
(7)
Calculated
average
mierotubule
length
(L)
tan
At xylem
thickenings
9*
9*
9*
9*
12"
12"
6
9
9
13
14
16
2
4
7
10
12
8
Pre-prophase
band
41"
21,
22*
33w
16
37
51
47
28
44
80
109
14w
14w
2011
12"
23w
23w
2011
5011
5011
7682
30**
7482
40*
90**
11
11
9
15
8
11
19
10
11
10
27
16
30
16
5
4
9
9
16
19
28
29
24
57
22
66
42
72
Interphase
4.0
2.8
1.7
1.7
1.9
3.4
mean 2.5 • 0.97
3.2
2.4
2.0
2.0
mean 2.4 • 0.57
4.2
5.5
2.7
2.9
1.7
1.8
1.9
2.4
3.8
1.9
5.1
2.5
3.8
2.9
mean 3.1 • 1.23
Each line in the table represents a different sequence of
serial sections. The first three columns of data give the
quantities needed to calculate the average microtubule
length (L) using the formula L = 2Na/T. Individual
and mean values for each category of microtubule array
are given in the last column.
Duration of fixation in glutaraldehyde:
*4h.
*8h.
w10-15 min.
112h.
82
** 17 h.
pre-prophase band (Fig. 8). Three bands of terminations are evident, all parallel to the plane of
sectioning and giving rise overall to significant
deviation from the poisson distribution. The microtubule terminations in the other sequences of
THE JOURNAL OF CELL BIOLOGY "VOLUME 7 7 , 1 9 7 8
TABLE II
Average Lengths of Cortical Microtubules in Zea and Impatiens Root Tip Cells
Species
No. of see.
tions in sequence
Type of microtubule array
Average no.
of microtubttles per see-
No. of terminations in se-
fion
quence
Zea mays
Pre-prophase band
Interphase
22
18
18
18
41
10
14
18
44
10
8
12
Impatiens balsamina
Interphase
30
19
24
11
25
20
10
12
10
Calculated average microtubule length
mean
mean
2.8
2.6
4.7
3.8
3.7 _ 1.05
4.4
5.4
6.9
5.6 --- 1.26
the table are one section thickness less than the
combined thickness of the sections in which the
microtubule was seen. The table includes only
complete microtubules up to 1.4 /zm in length,
although longer ones did occur in the longer
sequences. The advantage of expressing the numbers of complete microtubules in this way is that
sequences of similar total thickness can be compared. For example, in a comparison of conventional fixation versus fixation in polymerization
medium, no differences emerge. On the other
hand, some of the treatments to be described
below had marked effects on the length distribution of the microtubules.
Treatments Which Can Cause
1
1o
Depolymerization o f Microtubules
20
Number of section along sequence
1pm
FIGUR~ 7 An interphase array of microtubules in a
root fixed in glutaraldehyde dissolved in complete polymerization medium (see text).
sections of material fixed in this way were randomly distributed according to both poisson and
pattern analyses.
Complete Microtubules
Many of the sequences of serial sections included some complete microtubules. Table IV
lists the numbers of complete microtubules in
four length categories, expressed in each case as a
percentage of the total number of microtubules
present in the sequence. Because terminations
have been arbitrarily taken to occur halfway
through the section, the complete lengths cited in
Low TEMPERATURE: The cold treatments employed in this work did not lead to the
complete disappearance of microtubule arrays in
the root tip cells. 22.4% of all terminations found
in material fixed in the cold without any recovery
time were C-shaped- 13 times the frequency seen
in control tissues. In one sequence of sections,
two C-shaped terminations pointed in one direction and 10 in the other; in another sequence,
two pointed one way and 1 the other. The distribution of terminations was random in two sequences, but the one illustrated in Fig. 9 was
nonrandom according to the poisson and the
pattern analyses. Calculation of average lengths
(Table VI A) reveals a diminution to about half
that found in controls, and a rapid (15 min)
recovery to the normal range of lengths. The
proportion of short microtubules was only slightly
greater than in controls (Table IVB).
HARDHAM AND GUNNING CorticalMicrotubules in Plants
23
TABLE Ill
Average Lengths of Cortical Microtubules in Azolla after Fixation in a Microtubule-Polymerization Medium
Treatment
Type of microtubule array
No. of sections in sequence
Average no.
of microtubules per
section
No. of terminations in
sequence
19
16
6
8
6
14
8
8
21
39
17
46
11
10
15
8
18
19
11
19
14
98
4
4
4
2
4
9
7
35
Calculated average microtubule length
/.ml
Complete polymerization medium
Polymerizationmedium minus GTP
At xylem thickenings
Pre-prophase band
Interphase
Interphase
HIGH PRESSURE: The effect of high pressure is not straightforward. After a 30-min treatment at 6,000 lb/in2, the average length was 5.2
/zm (Table VIB), yet after 2 h recovery, the
value obtained was 1.5 /gin. Longer sequences
and higher pressures gave more consistent data,
15 min at either 14,000 or 16,000 lb/inz reduced
the average length to less than half of that seen in
controls (Table VI B).
A reconstruction of a 55-section sequence of a
pre-prophase band in a root fixed immediately
after 15 min at 14,000 lb/in2 is shown in Fig. 10.
Both poisson and pattern analyses indicate that
the terminations are nonrandomly distributed.
The proportion of short complete microtubules is
large (Table IV B), 28% being <0.35 /~m in
length, compared with 14 and 0% in this size
class in two 50-section sequences in controls. A
33-section sequence after 16,000 lb/in'~ treatment
also had augmented numbers of short microtubules. The proportion of C-shaped terminations
was as in controls, and only one sequence had
more than one present. In it, both C-shapes
pointed in the same direction.
COLCHICINE: Microtubules were present
in some Azolla root tip cells even after 2-5 h in 5
• 10 -'~ M colchicine solution. 3.6% of their
terminations were C-shaped (Table V). Two sequences had two C-shapes, and in each the two
24
8
13
10
13
12
8
18
22
mean
3.2
1.0
2.2
2.7
3.2
7.5
5.0
2.4
4.5
3.0
3.8 • 1.8
mean
2.2
2.1
6.2
2.7
3.3 --_ 1.95
6
7
4
5
C-shapes pointed in opposite directions. The percentage of short complete microtubules was
raised relative to the controls (Table IV B), and
the average microtubule lengths were approximately one-third to one-half those in untreated
roots, even after a 1-h recovery period (Table
VI C). In three of the six sequences, the microtubule terminations were nonrandomly distributed.
Treatment with D 2 0
Exposure of Azolla roots to D20 for 5 h
produced microtubule arrays which contained
large numbers of short microtubules (Table IV C
and Fig. 11). The average microtubule length was
approximately half that in controls (Table VII).
The frequency of C-shaped terminations was
0.9%. Only one sequence contained more than
one, and in it, all three C-shapes pointed in the
same direction. As illustrated in Fig. 4 B, short
microtubules appear in response to D20 treatment
amongst those that are found over developing
xylem thickenings; microtubules may in addition
be initiated between the thickenings, where their
frequency would normally be low (cf. Fig. 4 A).
After 18 h in D20, the average length was nearly
that in controls (Table VII), as was the length
distribution (Table IVC); the frequency of Cshaped terminations was zero (Table V).
THE JOURNAL OF CELL BIOLOGY - VOLUME 7 7 , 1 9 7 8
I I I L I ~ I ]0
~JJ
1
10
Number of section along sequence
0"5pm
FIGURE 8 A pre-prophase array in a root fixed in
glutaraldehyde dissolved in complete polymerization medium. Due to the complexity of this array, the mapping
process has in this case involved spacing the microtubules
along the plasma membrane and also straightening out
their undulations.
DISCUSSION
The major result reported here is that most cortical microtubules in plant cells are short relative to
the dimensions of the cell circumference. Microtubule arrays may extend over large or restricted
expanses of cell cortex, but the continuity of an
array is merely statistical, the constituent microtubules being present as overlapping units of
varying but limited length, on average one crosssectional profile in every 18 representing a terminating microtubule. Before discussing the implications of this observation, the validity of the
methods used must be examined.
Methods
No direct check on the efficacy of glutaraldehyde fixation with respect to cortical microtubules
has yet been devised. Attempts to use polarized
light microscopy as a nondestructive method to
assess microtubule densities in vivo, followed by
comparison with counts made after fLxation,
would not be suitable for the plant cell cortex
where any birefringence due to the microtubules
would be masked by that of the cell wall. Gluta-
raldehyde does, however, fix microtubules that
have been polymerized in vitro, and affords protection from alterations to their length (31). It
also preserves rings, sheets, and tubules in the
same proportions as observed in negatively
stained preparations of polymerizing mixtures,
and it prevents free tubulin from polymerizing
(25).
It was found that the duration of the glutaraldehyde fixation step can be altered between 10
min and 17 h with no marked effects upon microtubule length (Table I). Dispositions that might
be interpreted as being due to an originally intact
microtubule fracturing and the ends so formed
moving apart are rare in the reconstructions: one
example is in section 41 of Fig. 6 A. The existence
of cell-specific features in the back-to-back arrays
of Fig. 6 A and B also argues against major
mechanical damage during specimen processing.
Reports (30, 31) that relatively slow entry of
glutaraldehyde permits the buffer to alter the
degree of polymerization of tubulin within HeLa
cells appeared in the course of the present work.
In Azolla roots, unlike HeLa cells, the microtubule lengths are essentially the same whether the
glutaraldehyde is dissolved in phosphate buffer or
in polymerization medium with or without GTP.
If microtubules longer than those detected by
the tracking procedure were consistently found in
planes of section that show their longitudinal
aspect, then doubt would be cast on the methods
used here. In fact, the longest segment of cortical
microtubule found in a collection of several thousand micrographs of Azolla root tips measured
2.5 gm, despite the use of thick sections (0.2-0.5
p.m) and high voltage electron microscopy in part
of the work to search for longer profiles. All
papers on cortical microtubules listed in a recent
review (21) have been inspected, and again, no
microtubules longer than 1-2 /zm were found in
the published micrographs, with the exception of
two examples, at 3.9 and 3.6 /xm, in freezefractured pea root tip cells (36). Longer profiles
would be expected in freeze-fractured material,
as a fracture plane can follow a curved cell cortex
whereas a planar ultrathin section cannot. There
may, however, be an additional factor, for the
pea roots in question were soaked, without prefixation, for 1-10 days in 20% glycerol. Glycerol is
now known to support polymerization of tubulin
(49), binding to and allowing otherwise inactive
tubulin to assemble into microtubules (9). Thus,
there is no conflict between the longitudinal views
HARDHAM AND GUNNINO CorticalMicrotubules in Plants
25
TABLE IV
Percentagesof Complete Microtubules in Cortical Microtubule Arrays in Azolla Root Tip Cells
Treatment
A. Glutaraldehyde in
0.025 M phosphate
Type of microtubule Array
Interphase
0.42-0.7
0.77-1.05
1.12-1.4
2
13
2
14
0
3
4
0
0
2
7
6
2
10
12
4
4
2
0
2
2
8
8
2
7
5
40
18
27
8
5
5
4
0
4
0
0
5
Pre-prophase
Interphase
55
33
71
25
28
16
6
4
4
8
Interphase
19
29
39
31
18
48
19
26
27
31
15
0
4
11
16
14
11
4
8
3
29
50
56
34
35
38
194
66
27
23
13
25
23
13
14
Interphase
Xylem
Pre-prophase
Interphase
18 h
and the results obtained by tracking, and this,
together with the ability to detect interspecific
and intertreatment differences in microtubule
lengths, gives added confidence that the fixation
and tracking procedures used here are valid, at
least for the relatively simple arrays found in the
plant cell cortex.
Arrays o f Overlapping
Constituent Microtubules
Cortical arrays in plant cells are not the only
examples of systems composed of overlapping
microtubules. They occur in myogenic cells (59)
26
0-0.35
31
26
22
Interphase
4 h
2~ 15 min + 15 min
recovery
High pressure:
14,000 lb/in2
16,000 lb/in2
Colchicine
2h
2h
3h
5h
2 h + 1 h recovery
C. Deuterium oxide
5h
Complete microtubules, as % of total no.. in different
length classes (tLm):
52
39
51
24
27
32
52
109
38
95
58
33
14
41
14
Interphase
1.6~
Total no. of
microtubules
90
76
74
50
50
41
40
33
30
22
21
20
20
39
21
Pre-prophase
Interphase
Pre-prophase
Interphase
Pre-prophase
Glutaraldehyde in cornplete polymerization medium
B. Low temperature
0~ 15 min
No. of sections in sequence
T H E JOURNAL OF CELL BIOLOGY " VOLUME
77,
2
6
3
2
3
3
4
4
1
2
4
3
11
4
7
2
3
7
4
4
and in a variety of mitotic spindles: in nonkinetochore fibres in Haemanthus endosperm (23, 26);
in an insect spermatocyte (11), where < 1 % of
the tubules were traced and lengths in the range
of 1.0-5.5 /zm found; in mammalian cells in
tissue culture (32, 33); and in the fungi Thraustotheca (18) and Uromyces (20). The formulae used
in the present work can be applied to two of the
published sets of data. In myogenic cells (59),
data for one set of serial sections yield a calculated
average microtubule length of 27 /zm in a cell
- 1 5 0 /zm long, with the distribution of the observed terminations conforming to a poisson dis-
1978
tribution. In Uromyces (20) a 22-section sequence
contained complete microtubules between 1 and
15 section thicknesses in length, and the formula
gives an average length of 2/xm.
That cortical arrays of microtubules in plant
cells consist of overlapping microtubules provides
an explanation for the disparity in numbers that
has sometimes been recorded for different parts
of the same array, in wheat I and Azolla (present
work) root tip ceils, cells of Sphagnum "leaflets"
(46-48), in hypocotyl cells (45) and in the protonema of a fern (58). In both Azolla 2 and Sphagnum (48), the numbers are greater along the
younger wall. Such occurrences would not be
possible if the microtubules were in the form of
continuous hoops or spirals, and their implication
is that the density of the array alters across the
face of the cell, much as mapped in Figs. 5 and 8.
In at least some cases, particularly in pre-prophase
bands, it is very probable that the density differences arise because the development of the array
is incomplete or because it is breaking down
progressively from one side of the cell. The
lengths of the complete microtubules in the sequences of sections varied widely, and it is important to know whether there might be a proportion
of very long representatives, which might approach the dimensions of the "hoops" that are
referred to so often in the current literature (see
the introduction).
The points that are plotted in Fig. 12 represent,
for each sequence of serial sections, the number
of microtubules that passed right through the
sequence with neither termination included, normalized by expressing it as a fraction of the
average microtubule number in the array; this
fraction is plotted against the number of sections
in the sequence. The value of the fraction approaches unity as the number of sections per
sequence is reduced, and diminishes as the number of sections increases, reaching zero in two of
the longest sequences. Values for all fixation
regimes and all categories of microtubule array in
roots that had not been subjected to experimental
treatment were pooled before using the least
squares method to fit the exponential decay curve
shown in Fig. 12. The results may be interpreted
as showing an exponentially decreasing probability
of finding microtubules of length equal to or
O'Brien and Maynard. Personal communication.
z A. R. Hardham and B. E. S. Gunning. Manuscript in
preparation.
1
10
l
lO
Number of section
20
30
20
along sequence
30
11Jm
FIGURE 9 A pre-prophase band of microtubules in an
Azolla root fixed after 15 rain at 0~ This treatment
has resulted in an overall decrease in microtubule length.
greater than the combined section thicknesses, as
the number of sections per sequence is increased.
Extrapolation of the curve of best fit suggests that
if it were possible in practice to follow the whole
circumference of a 10-/zm diameter cell by serial
sectioning, the probability of finding a microtubule of length equal to the total circumference
would be 8 x 10 -~. The available data thus point
strongly to the absence of microtubulcs of length
comparable to the cell circumference.
Perturbations o f the Arrays
The data on treatments which in other systems
alter the stability of the tubulin:microtubule equilibrium in favor of either assembly or disassembly
are limited because the sample sizes obtained
through the use of serial sectioning are extremely
small. Further, particularly in the case of the
depolymerizing treatments, the sequences of sections are highly selected because the majority of
cells had lost a considerable proportion of their
cortical microtubules. It is not known whether
such losses might be selective, as in other systems
(3, 44). The maps (Figs. 9 and 10) and data
(Tables IVB and VI) presented here therefore
represent either surviving or recovering areas of
cell cortex, examined in the hope that the form of
the microtubules might give some insights into
the dynamics of the arrays.
The depolymerizing treatments fall into two
categories, though all three reduced the average
length (Table VI). Colchicine and high pressures
gave rise to arrays with greatly augmented proportions of short microtubules (Table IV) without,
HARDHAM AND GUNNING CorticalMicrotubules in Plants
27
I . I l t i l l k l l
9
t l l l t l l t l
po
2pL I L I I I L I I
20
~
40
5oI l l l l
40
50
t l i i ] l t L I i l l l l l l l i l
30
Number of section along sequence
1 prn
FI6URE 10 A pre-prophase band of microtubules in an Azolla root subjected to a pressure of 14,000
lb/in2 for 15 min, The array contains a large number of very short microtubules, 34% being <0.7/~m in
length, approximately three times the frequency seen in sequences of comparable length in untreated
roots.
TABLE V
Percentagesof C-Shaped Terminations
Treatment
Control
Polymerization medium
Low temperature
High pressure
Colchicine
5 h DzO
18 h D20
Total no. of
terminations
observed
Percent of Cshaped terminations
706
181
67
234
139
445
87
1.7
0
22.4
1.7
3.6
0.9
0
however, producing large increases in the proportion of C-shaped terminations (Table V). Low
temperature, on the other hand, had little effect
on the length distribution, but produced very high
frequencies of C-shapes. In this latter case, an
interpretation that is consistent with the data is
that the microtubules had (on average) shortened
through terminal disassembly, many of them
opening out to give the C-profile; the shortest
microtubules initially present had disappeared and
were replaced by remnants of longer ones, the
average length thus falling. It is not ruled out,
however, that some C-shapes might be generated
in the cold without accompanying disassembly.
The former pattern, induced by colchicine and
28
high pressure, is suggestive of fragmentation, presumably in addition to disassembly, or, alternatively, of early stages in recovery when the population contained short but growing microtubules.
In general, D20 is reported to stabilize microtubule arrays and enchance microtubule formation
from available pools of tubulin (22, 54). In plants,
no effects were found in wheat root tips (6) and
rye leaves (57), yet in Sphagnum leaf cells at the
same stage of the cell cycle there was a 25%
increase in the number of microtubules (48). For
pre-prophase bands, a marked increase in microtubule numbers was claimed for wheat roots (6),
yet in Sphagnum the band appeared more transient and indistinct (48). Observations of preprophase bands in untreated Azolla roots 3 show
great variation in microtubule numbers, presumably reflecting developmental changes, and the
map shown in Fig. 8 serves to emphasize that the
number of microtubules per section can be a poor
criterion for this type of experiment. More reliance can, however, be placed on observations of
microtubule length within a given array. A 5-h
treatment in D20 skewed the length distribution
strongly towards short microtubules as compared
3 B. E. S. Gunning, A. R. Hardham, and J. E. Hughes.
Manuscript in preparation.
THE JOURNALOF CELL BIOLOGY9 VOLUME 77, 1978
with controls (Table IV). This applied to all three
categories of array, and a likely interpretation is
that additional tubules were generated which,
being short, lowered the overall average length
(Table VII). There was no general diminution in
microtubule numbers throughout the root tip such
as was observed in the depolymerizing treatments.
A f t e r 18-h treatment, the length distribution and
the average length approximated to the control
situation, and in the absence of firm information
on total numbers per cell or per unit length of
wall, the only valid interpretation of this seeming
return to normality is that the response to D 2 0
does not involve enhancement of assembly onto
existing microtubules (which would have increased the average length), but rather the gener-
TAaLE VI
Average Lengths of Cortical Microtubules in Azolla Root Tip Cells alter Depolymerization Treatments
Treatment
Type of microtubule array
No, of sections in sequence
Average no. of
mka'otubules per
section
No. of terminations in sequence
Calculated average microtubule
length
gra
A. Low temperature
0~ 15 min
1.6~
Interphase
31
19
48
1.7
26
8
19
1.5
22
17
18
2.8
Interphase
Pre-prophase
14
14
21
74
8
95
5.2
1.5
Interphase
55
33
17
9
108
31
1.2
1.3
19
29
18
39
31
15
11
8
14
6
53
20
36
25
33
0.8
1.9
0.6
1.2
0.8
4h
2~ 15 rain + 15 min recovery
B. High pressure
6,000 lb/in 2, 30 min
6,000 lb/in 2, 30 min + 2 h
recovery
14,000 lb/inz, 15 min
16,000 lb/inz, 15 min
C. Colchicine
2h
2 h + 1 h recovery
3h
5h
Interphase
Number of section along sequence
1pm
FIGURE 11 An interphase array of microtubules in an Azolla root which had been immersed in D~O for
5 h. Dispersed throughout the array is a large number of short microtubules, the presence of which
lowers the calculated average microtubule length.
HARDHAM AND GUNNING CorticalMicrotubules in Plants
29
TABLE VII
Average Lengths ofCorticalMwrotubules~ AzollaRootTipCellsafierDeu~umOxide Treatment
Treatment
Type of microtubule array
No. of sections in
sequence
Average no. of microtubules per section
No. of terminations in sequence
Calculated average
microtubule length
11
13
29
50
13
10
56
8
13
13
12
35
34
7
16
15
50
9
21
16
46
10
15
20
13
16
15
17
38
293
21
14
96
24
7
14
11
19
22
0.7
1.7
1.7
1.3
0.8
2.1
1.3
2.2
2.7
1.9
3.1
3.2
3.4
tam
5 h D20
Xylem
Pre-prophase
Interphase
18 h D20
Pre-prophase
Interphase
ation of new ones.
Treatments which should favor polymerization
(presence of GTP, D20) diminished the proportion of terminations with C-shapes, whereas low
temperature and (to a lesser extent) colchicine
increased this proportion. It has been claimed
that the C-shape is associated with disassembly
(43), but it can also occur at growing terminations
in vivo (1) and in vitro (5) (for full discussion see
references 7, 23, and 26). C-shaped profiles were
not restricted to the ends of the microtubules,
even in untreated material, and low temperatures
enhanced the proportion of subterminal C-shapes.
Of the 325 complete microtubules that were
mapped, 15 had a visible manifestation of polarity, viz., a C-shaped profile at one end but not at
the other. No microtubules with C-shapes at both
ends were found. An obvious interpretation of
this morphological polarity is that the microtubules grow or disassemble preferentially at one
end, as has been observed in vitro (see references
8 and 51). It is therefore of great interest that Cshapes are patterned in the microtubule maps.
When serial sections are examined, the C-shapes
point in the same direction. The observed numbers of C-shapes were as follows (expressed in
the form of the number of C-shapes pointing one
way:number pointing the opposite way) 0:2, 0:4,
0:6 (controls); 0:3 (5 h, DzO); 0:2 (high pressure); 2:10, 1:2 (low temperature); and 1:1, 1:1
(colchicine). The observations on colchicine-affected arrays are few, but suggest that the microtubules in them, although still polar as individuals,
do not share a common polarity. In the controls
30
and DeO treatment, at least those microtubules
that have C-shaped terminations share a common
morphological directionality. It is not known
whether this sharing extends to the majority of
the microtubules (those lacking C-shaped terminations), nor is it known whether all faces, or
indeed all parts of faces, of a cell share the same
directionality. The cell cortex could be subdivided
into domains that differ from one another but
within which directionality is shared.
Development o f Cortical Arrays
In undertaking the present work, one expectation was that evidence for microtubule-organizing
structures might emerge. In fact, the great majority of microtubule terminations are not patterned,
with the exception of some clustering (e.g. Fig. 2
in reference 17). Precedents for linear microtubule-organizing centers (MTOCs) which generate
microtubules at right angles to their long axis do
exist in the algae (2, 4), but no ultrastructural
evidence for any equivalent of these rhizoplast
fibers was found here. The lines of terminations
may well be significant, particularly the two that
lay back-to-back in neighboring cells, in or close
to the direction of cell elongation (Fig. 6). They
can, however, hardly be indicative of MTOCs
unless these are so few in number that the sample
sequences of sections rarely included them, or
unless some form of propagation and turnover of
the population of microtubules obliterates a patterned origin that is conspicuous only transiently
in an earlier stage of the cell cycle. It must be
emphasized that the sample sizes are not only
T H E JOURNAL OF CELL BIOLOGY" VOLUME 7 7 , 1 9 7 8
small relative to the surface area of the cell, but
also selective: because of the practical problems
of tracking microtubules, they did not penetrate
into any cell corners, which therefore remain as
possible sites for MTOCs.
The majority of cortical microtubules, short or
long, lie parallel to and at varying distances from
the plasma membrane, this being especially evident in highly stacked pre-prophase bands. Detailed examination of the serial sections shows
that microtubules, even very short ones, do not
bend towards and terminate "in or on the plasma
membrane," as suggested for Allium root tip preprophase bands (37). The source and mechanism
of placement of any initiator molecules that may
be required (see references 5 and 51) are therefore unknown. It should be pointed out that the
microtubules are not necessarily assembled where
they are seen. Neither the literature nor the
present results preclude the concept of an assem-
1"0
0.9
0"8'
9
0"6 ~
/A&
0"5"
0-2 84
o-I.
9
D.
D
Number of sections
The fraction of microtubules in an array
12
which pass right through the sequence of sections,
plotted against the number of sections in the sequence.
Squares, interphase arrays; circles, pre-prophase bands;
triangles, arrays overlyingxylem thickenings. Open symbols, conventional fixation; dosed symbols, fixation in
polymerization medium + GTP; half-closed squares,
fixation in polymerization medium without GTP. The
curve was fitted by the least squares method and represents the relationship y = e-~176 where y is the
ordinate and s is the number of sections in a sequence.
FIGURE
bly process that is followed by mobility of microtubules within the arrays. Once initiated, the
ultimate length of the microtubules could be governed by the balance between the numbers of
free tubulin molecules and the numbers of competing assembly sites, as demonstrated in vitro in
experiments where a given quantity of tubulin
can be used to make many short microtubules or
fewer longer ones (5, 50). Further experiments
with D20, which, as described above, seem to
induce additional microtubules rather than elongate existing ones, might be helpful in this context9
Some estimates of microtubule growth rates
are available. In plant material, microtubule
growth keeps pace with the 1-2 /xm min 1 elongation rate of root hairs (35). Values up to 1.4
/xm min ~ for in vitro growth are reported, dependent upon the tubulin concentration and the
concentration of polyanion (5). The rate of 1.5
ttm min -1 claimed for reassembling mitotic spindles (12) may be an overestimate in that it rests
on the assumption that the spindle fibers, which
were measured by optical microscopy, consist of
continuous rather than overlapping microtubules,
with single rather than multiple growing points.
The Azolla roots used here were elongating, and
unpublished data on the interpolation of additional microtubules along the cell walls suggest
that their growth rates approach the above values.
It is not known whether, or how, turnover of the
components of cortical microtubule arrays is accomplished. At one extreme, entire microtubules
could disassemble and be replaced at the growth
rates suggested above, in which case the lifetime
of a 2-4 ~m long microtubule might be only a
few minutes. The low frequency (1.7%) of Cshaped terminations militates against this view.
Alternatively, individual microtubules, once
formed, might be stabilized by becoming crossbridged to adjacent structures.
Bridges occur between adjacent microtubules
and between microtubules and membranes in
Azolla, as in other plant cells (21). The present
observation of periodic cross bridging in arrays of
plant cortical microtubules is of potential significance with respect to possible functions, be they
in microtubule assembly, in stabilizing and orienting the microtubules, in stabilizing domains or
molecules in or on the plasma membrane, or in
mediating motility phenomena. We are grateful
to Dr. P. B. Green (Stanford University, Stanford, Calif.) for pointing out to us that, in arrays
of overlapping component microtubules such as
HARDHAM AND GUNNING Cortical Microtubules in Plants
31
we have detected, mechanochemical interactions
which tend to maximize the extent of overlap will
influence the overall orientation. In the case of
the side walls of approximately cyclindrical cells
such as examined here, maximization of overlaps
would hold the constituent microtubules under
tension, the array as a whole becoming predominantly transversely oriented.
Microtubules and Micro fibrils
Much of the preceding discussion has centered
on the properties of the extensive microtubule
arrays that lie against the longitudinal walls of
elongating Azolla root tip cells, where the microtubules and the cellulose microfibrils4 share a
predominantly transverse orientation, just as in
other species. However, deposition of cellulose
can also be precisely localized in small areas of
the cell surface, as at pits (42) or the thickenings
of stomatal pores (38). Whereas it is difficult to
envisage how arrays of long "hooplike" microtubules could be locally differentiated in order to
participate in such events, the present observation
that cortical arrays consist of short overlapping
components does at least allow the concept of
microtubule participation to be applied to both
large and small expanses of wall: the only difference in principle is the degree to which the array
of overlapping microtubules is propagated around
the cell cortex. At both pits (42) and stomatal
pores (38) the microtubules and wall microfibrils
are co-aligned, which confronts us once again
with the major problems of how microtubule
arrays become localized and oriented. It may be
that the systems that control the formation of
extensive as compared to spatially restricted arrays
are considerably different.
The observations presented here have no direct
bearing on the possible role of cortical microtubules in orienting the deposition of cellulose microfibrils. However, the view that the cortical
microtubules represent "relatively rigid tracks
along which other cellular components (such as
cellulose synthetase complexes projecting inwards
through the plasma membrane to the microtubules) might move" (19) is now unattractive because the "tracks" are seen to be too short. Guidance of cellulose synthetase complexes by means
of microtubule movements generated by intertu4 Polarized light microscope observations by Dr. P. B.
Green, Stanford University, personal communication.
32
bule sliding or polarized assembly-disassembly is
not ruled out, nor is indirect guidance by the production of shearing forces in the membrane by
motility-generating molecules associated with the
cortical microtubules (21). Indeed, the latter suggestions receive a measure of support from the
detection of polarity in the cortical arrays.
We thank Drs. D. J. Carr, P. B. Green, P. K. Hepler,
and E. H. Newcomb for valuable discussions in the
course of the work and the preparation of the manuscript. The hospitality and help given to B. E. S.
Gunning in the High Voltage Electron Microscopy Laboratory in the Department of Molecular, Cellular and
Developmental Biology in the University of Colorado,
Boulder, is gratefully acknowledged.
Received for publication 8 September 1977, and in
revised form 2 December 1977.
REFERENCES
1. BEHNKE, O. 1967. Incomplete microtubules ob-
2.
3.
4.
5.
6.
7.
8.
9.
10.
THE JOURNALOF CELL BIOLOGY9 VOLUME77, 1978
served in mammalian blood platelets during microtubule polymerization.J. Cell Biol. 34:697-701.
Boucg, G. B., and D. L. BRown. 1973. Microtubule biogenesis and cell shape in Ochromonas. I.
The distribution of cytoplasmic and mitotic microtubules. J. Cell Biol. 56:340-359.
BRINKLEY,B. R., and J. C~a~xwmGm,JR. 1975.
Cold-labile and cold-stable microtubules in the mitotic spindle of mammalian cells. In The Biology of
Cytoplasmic Microtubules. D. Soifer, editor. Ann.
N. Y. Acad. Sci. 253:428-439.
BRows, D. L., and G. B. BOUCK.1973. Microtubule biogenesis and cell shape in Ochromonas. II.
The role of nucleating sites in shape development.
J. Cell Biol. 56:360-378.
BRYAN, J. 1976. A quantitative analysis of microtubule elongation. J. Cell Biol. 71:749-767.
BURGESS, J., and D. H. NORTnCOTE.1969. Action
of colchicineand heavy water on the polymerization
of microtubules in wheat root meristems. J. Cell
Sci. 5:433-451.
COHEN, W. D., and T. GO'rrLIEa. 1971. C-microtubules in isolated mitotic spindles. J. Cell Sci.
9:603-619.
DENTLER, W. L., S. GRANETr, G. B. WmaxN,
and J. L. ROSENBAUM. 1974. Directionality of
brain microtubule assembly in vitro. Proc. Natl.
Acad. Sci. U. S. A. 71:1710-1714.
DETRICH,H. W., III, S. A. BERKOWITZ,H. KIM,
and R. C. WILLiAmS,JR. 1976. Binding of glycerol
by microtubule protein. Biochem. Biophys. Res.
Commun. 68:961-968.
FISHER, R. A. 1948. Statistical Methods for Research Workers. Oliver and Boyd Ltd., Edinburgh.
11. FUr;E, H. 1974. The arrangement of microtubules
and the attachment of chromosomes to the spindle
during anaphase of Tipulid spermatocytes. Chromosoma (Bed.). 45:245-260.
12. GOODE, D. 1973. Kinetics of microtubule formation
after cold disaggregation of the mitotic apparatus.
J. Mol. Biol, 80:531-538.
13. GREEN, P. B., R. O. ERICKSON, and B. A. RICHMONO. 1970. On the physical basis of cell morphogenesis. Ann. N. Y. Acad. Sci. 175:712-731.
14. GREm-Srarrn, P. 1952. The use of random and
contiguous quadrats in the study of the structure of
plant communities. Ann. But. (Lond.). 16:293316.
15. GUNNING, B. E. S., and A. R. HAaDrlAM. 1977.
Estimation of the average section thickness in ribbons of ultrathin sections. J. Microsc. 109:337341.
16. GUNNING, B. E. S., A. R. HAROrlAM, and J. E.
HUGHES. 1976. Morphogenesis and microtubules
in Azolla root apices. J. Cell Biol. 70(2, Pt. 2):
56a. (Abstr.).
17. HAROHAM, A. R., and B. E. S. GUNNING. 1977.
The length and disposition of cortical microtubules
in plant cells fixed in glutaraldehyde-osmium tetroxide. Planta. ( Berl.). 134:201-203.
18. HEATH, I. B. 1974. Mitosis in the fungus Thraustotheca clavata. J. Cell Biol. 60:204-220.
19. HEATH, I. B. 1974. A unified hypothesis for the
role of membrane bound enzyme complexes and
microtubules in plant cell wall synthesis. J. Theor.
Biol. 45:445-449.
20. HEATH, I. B., and M. C. HEATH, 1976. Ultrastructure of mitosis in the cowpea rust fungus Uromyces
phaseoli var. vignae. J. Cell Biol. 70:592-608.
21. HEPLER, P. K., and B. A. PALEVrrz. 1974. Microtubules and microfilaments. Annu. Rev. Plant Physiol. 25"309-362.
22. INOU~, S., and H. RrrrER, JR. 1975. Dynamics of
mitotic spindle organization and function. In Molecules and Cell Movement. S. Inou6 and R. E.
Stephens, editors. Raven Press, New York. 3-30.
23. JENSEN, C., and A. BALER, 1973. Spindle dynamics
and arrangement of microtubules. Chromosoma
( Berl.) . 44:73-89.
24. KERSFIAW,K. A. 1973. Quantitative and Dynamic
Plant Ecology. Arnold (Edward) (Publishers) Ltd.,
London.
25. KIRSCHNER, M. W., L. S, HONIG, and R. C.
WILLIAMS. 1975. Quantitative electron microscopy
of microtubule assembly in vitro. J. Mol. Biol.
99:263-276.
26. LAtaaERT, A.-M., and A. S. BALER. 1972. Dynamics of spindle fibres and microtubules during ariaphase and phragmoplast formation. Chromosoma
(Bed.). 39:101-144.
27. LEDBETrER, M. C. 1967. The disposition of microtubules in plant cells during interphase and mitosis.
Syrup. Int. Soc. Cell Biol. 6:55-70.
28. LEDBETrER, M. C., and K. R. PORTER. 1963. A
"microtubule" in plant cell fine structure. J. Cell
Biol. 19:239-250.
29. LEOaETIXR, M. C., and K. R. PORaXR. 1970.
Introduction to the Fine Structure of Plant Cells.
Springer-Verlag, Berlin.
30. LuFrm, R. B,, P. N. MCMILLAN, J. A. WEATHERBEE, and R. R. WEImN6. 1976. Increased visualization of microtubules by an improved fixation
procedure. J. Cell Biol. 70(2, Pt. 2):228a. (Abstr.).
31. Lor'n~, R. B., P. N. McMILLAN, J. A. WEATHERnEE, and R. R. WEImN6. 1977. Increased visualization of microtubules by an improved fixation
procedure. J. Histochem. Cytochem. 25:175-187.
32. Mclrerosn, J. R., W. Z. CANOE, and J. A. SNVDER.
1975. Structure and physiology of the mammalian
mitotic spindle. In Molecules and Cell Movement.
S. Inou6 and R. E. Stephens, editors. Raven Press,
New York. 31-76.
33. MClNTOSH, J. R., Z. CANDE,J. SNYDER,and K.
VANDERSLICE. 1975. In The Biology of Cytoplasmic Microtubules. D. Soifer, editor. Ann. N. Y.
Acad. Sci. 253:407-427.
34. NEWCOMB,E. H. 1969. Plant microtubules. Annu.
Rev. Plant Physiol. 20:253-288.
35. NEwcoma, E. H., and H. T. BONNErr, JR. 1965.
Cytoplasmic microtubule and wall microfibril orientation in root hairs of radish. J. Cell Biol. 27:575589.
36. NORTHCOTE, D. H., and D. R. LEVaS. 1968.
Freeze-etched surfaces of membranes and organelles in the ceils of pea root tips. J. Cell Sci. 3:199206.
37. PACKARD,M. J., and S. M. STACK. 1976. The preprophase band: possible involvement in the formation of the cell wall. J. CellSci. 22'403-411.
38. PALEVITZ, B. A., and P. K. HEPLER. 1976. Cellulose microfibril orientation and cell shaping in developing guard cells of Allium: the role of microtubules and ion accumulation. Planta (Berl.). 132:7193.
39. PICKETT-HEAPS, J. D. 1966. Incorporation of radioactivity into wheat xylem wails. Planta (Bed.).
71:1-14.
40. PICKEaI-HEAPS, J. D. 1967. The effects of colchicine on the ultrastructure of dividing plant cells,
xylem wall differentiation and distribution of cytoplasmic microtubules. Dev. Biol. 15:206-236.
41. PICKETT-HEAPS, J. D. 1974. Plant microtubules.
In Dynamic Aspects of Plant Ultrastructure. A. W.
Robards, editor. McGraw-H/ll, New York. 219255.
42. RoB~os, A. W., and P. G. HUMrHE~ON. 1967.
Microtubules and angiosperm bordered pit formation. Planta (Bed.). 77:233-238.
43. RUTH, L. E., and Y. SmGENAKA. 1970. Microtubules in the Heliozoan axopodium. II. Rapid deg-
HARDHAM AND GUNNING Cortical Microtubules in Plants
33
44.
45.
46.
47.
48.
49.
50.
34
radation by cupric and nicketous ions. J. Ultrastruct.
Res. 31:356-374.
SALMON, E. D., D. GOODE, T. K. MAUGEL, and
D. B. BONAR. 1976. Pressure-induced depolymerization of spindle microtubules. III. Differential
stability in HeLa cells. J. Cell Biol. 69:443-454.
SAWnNEY, V. K., and L. M. SRIVASTAVA. 1975.
Wall fibrils and microtubules in normal and gibberellic-acid-induced growth of lettuce hypocotyl cells.
Can. J. Bot. 53:824-835.
SCHNEPF, E. 1973. Mikrotubulus-Anordnung undUmordnung, Wandbildung und Zellmorphogenese
in jungen Sphagnum-Bl~ttchen. Protoplasma
(Bed.). 78:145-173.
SCHNEPF, E. 1974. Microtubules and cell wall
formation. Port. Acta Biol. Set. A. 14:451-462.
SCHNEPF, E,, G. DEICHGRg~ER, and N. LJUeE~Id. 1976. The effects of colchicine, ethionine and
deuterium oxide on microtubules in young Sphagnum leaflets. A quantitative study. Cytobiologie.
13:341-353.
SHELANSKI,M. L., F. GASrdN, anti C. R. CANTOR.
1973. Microtubule assembly in the absence of
added nucleotides. Proc. Natl. Acad. Sci. U. S. A.
70:765-768.
SLOBODA,R. D., W. L. DENTLER, and J. L. ROSENBAUM. 1976. Microtubule-associated proteins and
the stimulation of tubulin assembly /n vitro. Biochemistry. 15:4497-4505.
51. SNYDER, J. A., and J. R. Mclmosn. 1976. Bio-
52.
53.
54.
55.
56.
57.
58.
59.
THE JOURNAL OF CELL BIOLOGY " VOLUME 77, 1978
chemistry and physiology of microtubules. Annu.
Rev. Biochem. 45:699-720.
SrURR, A. R. 1969. A low-viscosity epoxy resin
embedding medium for electron microscopy. J.
Ultrastruct. Res. 26:31-43.
SRIVASTAVA,L. M., and A. P. SrSGH. 1972. Certain aspects of xylem differentiation in corn. Can.
J. Bot. 50:1795-1804.
STErHENS, R. E. 1973. A thermodynamic analysis
of mitotic spindle equilibrium at active metaphase.
J. Cell Biol. 57:133-147.
TnOMVSON, H. R. 1958. The statistical study of
plant distribution patterns using a grid of quadrats.
Aust. J. Bot. 6:322-342.
VENABLE, J. H., and R. COGGESHALL. 1965. A
simplified lead citrate stain for use in electron
microscopy. J. Cell. Biol. 25:407-408.
WAaER, J., and W. S. SAgAI. 1975. Further studies
of the ultrastructure of DzO grown winter rye.
Protoplasma. (Bed,). 84:273-281.
WADA, M., and T. P. O'BPaEN. 1975. Observations
on the structure of the protonema of Adiantum
capillus-veneris L. undergoing cell division following
white-light irradiation. Planta (Bed.). 126:213-227.
WAaREN, R. H. 1974. Microtubular organization
in elongating myogenic cells. J. Cell Biol. 63:550566.